What is the risk that you cannot sell?
Every measure so far — VaR, Expected Shortfall, stress losses — shares a silent assumption so basic it usually goes unstated: that you can transact at the price on the screen.
Mark a position at $50 and your risk model treats it as $50. But a mark is a quote, and a quote is an offer to trade a certain size. Liquidity risk is the gap between the price you are marked at and the price you could actually achieve, at your size, in the time you have.
Two different liquidity risks
Market (asset) liquidity risk — you cannot sell the asset quickly without moving the price against yourself. The bid-ask spread widens, the depth at each price level thins, and a size that traded invisibly last month now moves the market when you attempt it.
Funding liquidity risk — you cannot obtain the cash to meet obligations: a margin call, a redemption request, a maturing loan you assumed you could roll.
They are distinct, and they feed each other viciously. A funding shortfall forces you to sell into a market whose liquidity is evaporating; the resulting price impact deepens losses, which triggers further margin calls. This spiral is the mechanism behind most rapid institutional failures — including LTCM in 1998, whose positions were defensible on paper but could not be exited at anything near their marks.
Putting a number on it — days to liquidate
The simplest useful metric asks how long an orderly exit would take.
Days to liquidate = Position size ÷ (Participation rate × Average daily volume)
The participation rate is the fraction of a day's trading you are willing to be — go much above 15–25% and you become the market rather than a participant in it.
Worked example: you hold 5,000,000 shares. The stock's average daily volume is 2,000,000 shares. At a 20% participation rate:
5,000,000 ÷ (0.20 × 2,000,000) = 5,000,000 ÷ 400,000 = 12.5 trading days
Two and a half weeks of continuous selling to exit one position — assuming volume holds up. It rarely does in a crisis. Now compare that to a 1-day VaR figure. The mismatch is the point: your risk horizon is not one day if your exit horizon is twelve.
Liquidity-adjusted VaR
A common adjustment adds the cost of crossing the spread:
Liquidity-adjusted VaR = VaR + (½ × bid-ask spread × position value)
On a $10,000,000 position with a normal spread of 5 basis points, that adds ½ × 0.0005 × $10,000,000 = $2,500. Negligible.
Now shock the spread. In stressed markets, spreads on less-liquid instruments have historically widened many times over. At 80 basis points: ½ × 0.008 × $10,000,000 = $40,000 — sixteen times larger, on the same position, purely from the cost of getting out. Liquidity is cheap until it is the only thing that matters.
Where it has bitten
The recurring shape is a maturity mismatch: a vehicle promising fast redemption while holding assets that cannot be sold fast.
UK open-ended property funds suspended redemptions after the 2016 referendum and again in March 2020 — daily-dealing funds holding buildings, which do not trade daily. In September 2008 a large money-market fund "broke the buck," triggering redemptions across a sector investors had treated as cash-equivalent. In March 2020 even the US Treasury market — the deepest in the world — showed severe strain during the scramble for cash, until the central bank intervened.
The pattern is not exotic assets behaving badly. It is the promise of liquidity outrunning the liquidity of what backs it.
A correct hedge can still force a sale. In 2022 UK defined-benefit pension schemes held long gilts and swaps, often leveraged, to match liabilities. When 30-year gilt yields rose by more than a percentage point in the days after the 23 September 2022 fiscal statement, the hedges lost value as designed, since liabilities fell by more, but the losses were due in cash as collateral within days, and the asset that raised cash fastest was the gilt itself. Gilt sales pushed yields higher and raised the next call. The Bank of England began buying long gilts on 28 September 2022 to break the loop. Funding improved; liquidity nearly failed. Both are true, and only the first appears in a model marking assets at screen prices.
In the data
The measurable half of liquidity is how much changes hands each day. Average daily turnover in dollars over the last fifty sessions, for a large company, Apple, and a thin one, Escalade, a sporting-goods maker:
Put a position against a fifth of either figure and you have days to liquidate. What no daily price record shows is the bid-ask spread you would pay on the way out, least of all the stressed spread the formula above needs, so that half of liquidity risk has to be estimated rather than read.
Try it now
- Compute days-to-liquidate for a $5,000,000 position in each of the two names above at a 20% participation rate, and compare. To try a thin name of your own choosing, open it in the Terminal and read its daily volume and price there.
Open ESCA.US in the EODHD Terminal
- Now watch the number move. The chart below plots daily volume rather than price for a small-cap basket — find a stressed stretch and a calm one on it, then recompute days-to-liquidate from each. Note whether volume rose or fell exactly when you would most want it.
- Estimate liquidity-adjusted VaR for one position using a normal spread and then a spread ten times wider. Write one sentence about what that gap means for a one-day risk measure. A description of the exposure — never a statement about what level is acceptable.
Build it yourself
Build a watchlist that shows volume beside price, so thinness is visible before it matters. Your tickers, in one call